Structure

Introduction

general aspects of parallel computing and GPGPU fundamentals;

Second part

practical details (CUDA, OpenCL).

Stop me if I’m going to fast.

Interrupt me if I’m not being clear. Ask earlier rather than later.

(More detailed) Overview

• C and C++ programming basics
• Fundamental concepts of parallel programming
• History of graphic cards and GPGPU
• GPGPU programming basics; introduction to CUDA
• High-level CUDA programming: the CUDA runtime
• Benchmarking, optimization and debugging
• Low-level CUDA programming: the CUDA driver interface
• OpenCL basics and heterogeneous GPGPU programming
• Introduction to multi-GPU

Choices made

CUDA

by NVIDIA, for NVIDIA GPUs, didactically more appropriate to introduce some of the concepts

OpenCL

cross-platform (Mac OS X, Linux, Windows, Android), multi-vendor (NVIDIA, AMD, Intel, IBM, Altera, ARM, etc) standard for parallel computing. Supports CPUs, GPUs, DSPs, FPGAs. The same program can run on any supported hardware. Solution used by a multitude of programs and libraries: OpenCV, MATLAB, PhotoShop, LibreOffice, etc

Table of contents

• Introduction to parallel computing
• Fundamentals of GPGPU
• CUDA
• OpenCL